Automated video advertisement content analysis system using Sentence Transformers and cosine similarity for yes/no question evaluation. Features text embedding with all-mpnet-base-v2, batch processing, vector indexing, and performance evaluation against human-coded ground truth data.
natural-language-processing text-classification query-engine semantic-similarity cosine-similarity text-analytics batch-processing binary-classification content-analysis performance-evaluation text-embedding document-processing advertisement-analysis sentence-transformers vector-indexing video-ad-evaluation mpnet-model ground-truth-evaluation
-
Updated
Oct 23, 2024 - Jupyter Notebook